Your Search on the Marketplace is Getting More Relevant

[#1]

Atlassian has upgraded the search functionality on the Atlassian Marketplace, and if you're a partner or a frequent user, you'll want to be aware of this update.

Whether you're browsing for apps on the Marketplace website or searching directly from Jira, Confluence, or Bitbucket, finding the right app just got easier.

Search is now more accurate and relevant

The fundamental shift is in how the Marketplace understands what people are looking for.

Instead of asking, “Do these words match?”, search also asks, “What is this person trying to do?” – the new system interprets user intent and context.

What does this mean in practical terms? When someone searches for an app, they're now more likely to see results that actually solve their problem. Not just results that happen to contain their search terms.

The system understands relationships between concepts, recognizes business needs, and surfaces apps that are genuinely relevant to what users are trying to accomplish.

What you need to know about this change

Behind the scenes, Atlassian has migrated their search technology from Algolia to OpenSearch.

This rollout is happening in phases through January 31, 2026, and it's already changing how search results are ranked and displayed.

The old system relied mainly on keyword matching. The new one goes further by understanding what you're actually trying to accomplish. It uses natural language processing to interpret your intent, not just match your words.

What does that mean in practice:

OpenSearch considers not only whether keywords appear in an app's listing, but also whether that app genuinely solves the problem you're describing.

It weighs factors like:

• How well an app matches your search intent

• How engaged other users are with it

• How relevant it is to what you're looking for

For partners, this might mean your app's position in search results shifts a bit. But that's actually a good thing – it means customers who are genuinely looking for what you offer are more likely to find you.

Use your natural language to search on Atlassian Marketplace

This is where things get interesting. You no longer need to guess the exact keywords an app developer used in their listing. Instead, you can search the way you'd naturally describe your problem.

Let's say you need something to help your team track time. Previously, you'd need to search for specific terms like "resource planning" or "resource management" and hope the app you needed used those exact words.

Now, you can type something like "best tools for resource planning" and get relevant results.

The search engine understands context and relationships between words. It recognizes business needs and user roles. Then it surfaces apps that match those needs – even when you're not using the exact terminology the developer used.

So... does this change the way you search for apps?

Not really – but it does make searching better.

You don't need to learn a new way to search or use special tricks to find what you need. Just type naturally, the way you'd describe the problem to a colleague.

What you will notice is that you'll find relevant apps faster and with less trial and error. Instead of scrolling through pages of results trying to decipher which app does what, you'll see options that more closely match what you're looking for right at the top.

Type a full sentence if that feels natural – the search engine is built to handle it. Describe your use case, mention your workflow, or simply state the problem you're trying to solve. The more clearly you express what you need, the better the results you'll get.

What's powering these improvements

If you're curious about how this works under the hood, three main factors determine which apps appear in your search results and in what order on Atlassian Marketplace.

• Keyword relevance. This is still the foundation. The search looks at how well the terms you use match the text in an app's listing – especially the app name, tagline, and description.

• Semantic matching. This is what allows the search to understand meaning, not just words. The system generates what's called a vector representation of your search query and compares it to similar representations of app listings.

• Engagement signals. The search considers how other users interact with apps – things like the number of reviews, average ratings, install counts, and whether the partner provides support details. Apps that have proven useful to other customers get a relevance boost, which helps surface quality options.

These three factors work together to balance relevance with popularity, ensuring you see apps that both match what you're searching for and have a track record of helping users like you.

For Marketplace partners: What this means for your app

If you're a Marketplace partner, you might see your app's search position change as this improvement rolls out. This isn't necessarily good or bad – it's about better matching apps to customer intent.

What you should do to maintain or improve visibility:

• Focus on clear communication over keyword stuffing. Explain what problems your app solves using natural, conversational language.

• Build engagement signals. Encourage satisfied customers to leave reviews and ratings – these directly influence your search ranking.

• Show, don't just tell. Add screenshots, videos, or demos that help potential customers quickly understand your app's value.

• Provide support details. Include clear support information in your listing. This builds trust and contributes to your relevance score.

The goal here isn't to game the system, but to create listings that genuinely help customers understand whether your app is right for them.

What's next

The rollout is happening gradually, with 20% of users already using it, 50% by mid-January, and 100% by the end of January 2026. Atlassian is monitoring search quality metrics at each milestone to ensure everything is working smoothly.

For most users, this will be a seamless transition. You'll notice that search results feel more relevant and that finding the right app takes less effort. And that's exactly the point – making the Marketplace experience more intuitive and helpful for everyone.

[Calogero Bonasia]

The shift to semantic search via OpenSearch is technically sound, but the reliance on engagement signals as ranking factors introduces structural bias favoring incumbents. Vector-based semantic matching can identify intent, but if established apps dominate results due to historical review volume, the system risks perpetuating existing market concentration rather than surfacing objectively better solutions.

For this to genuinely improve discovery, Atlassian would need to time-weight engagement signals or introduce recency factors that allow newer apps to compete on semantic relevance alone. Otherwise, 'understanding what users are trying to accomplish' becomes secondary to 'showing what has historically been popular'—which isn't the same thing.

[Liam - DevSamurai]

Thanks for the feedback! @Calogero Bonasia 

As a Marketplace Partner, we share your concern.

We keep optimizing our listings to perform as best as possible in the current system, but agree that stronger recency weighting or more emphasis on pure semantic relevance would make discovery fairer and more innovative.

[Yana Misiura_BrizoIT]

that's a significant change. Thanks for sharing!
@Anna Vozniuk 

[Shivam Sharma - Optimizory]

That's actually a really good write!!
Kudos to you @Liam - DevSamurai for sharing this, for marketplace partners, I believe it's a game changing thing, because right now, vendors play with keywords to list them on top, but now who's the real problem solver getting privileged.